G Model
VETPAR-7722; No. of Pages 9 ARTICLE IN PRESS
Veterinary Parasitology xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Veterinary Parasitology
jou rnal homepage: www.elsevier.com/locate/vetpar
Taxonomy and molecular epidemiology of Echinococcus granulosus
sensu lato
∗
T. Romig , D. Ebi, M. Wassermann
Universität Hohenheim, FG Parasitologie 220 B, 70599 Stuttgart, Germany
a r t i c l e i n f o a b s t r a c t
Keywords: Echinococcus granulosus, formerly regarded as a single species with a high genotypic and phenotypic
Echinococcus granulosus
diversity, is now recognised as an assemblage of cryptic species, which differ considerably in morphol-
Taxonomy
ogy, development, host specificity (including infectivity/pathogenicity for humans) and other aspects.
Nomenclature
This diversity is reflected in the mitochondrial and nuclear genomes and has led to the construction of
Molecular epidemiology
phylogenetic trees and hypotheses on the origin and geographic dispersal of various taxa. Based on phe-
notypic characters and gene sequences, E. granulosus (sensu lato) has by now been subdivided into E.
granulosus sensu stricto (including the formerly identified genotypic variants G1-3), Echinococcus felidis
(the former ‘lion strain’), Echinococcus equinus (the ‘horse strain’, genotype G4), Echinococcus ortleppi (the
‘cattle strain’, genotype G5) and Echinococcus canadensis. The latter species, as recognised here, shows
the highest diversity and is composed of the ‘camel strain’, genotype G6, the ‘pig strain’, genotype G7,
and two ‘cervid strains’, genotypes G8 and G10. There is debate whether the closely related G6 and
G7 should be placed in a separate species, but more morphological and biological data are needed to
support or reject this view. In this classification, the application of rules for zoological nomenclature
led to the resurrection of old species names, which had before been synonymised with E. granulosus.
This nomenclatural subdivision of the agents of cystic echinococcosis (CE) may appear inconvenient for
practical applications, especially because molecular tools are needed for identification of the cyst stage,
and because retrospective data on ‘E. granulosus’ are now difficult to interpret without examination of
voucher specimens. However, the increased awareness for the diversity of CE agents – now emphasised
by species names rather than genotype numbers – has led to a large number of recent studies on this issue
and a rapid increase of knowledge on geographical spread, host range and impact on human health of the
various species. E. granulosus s.s., often transmitted by sheep, is now clearly identified as the principal
CE agent affecting humans. Contrary to previous assumptions, genotypes G6/7 of E. canadensis readily
infect humans, although CE incidences are rather low where E. canadensis predominates. Sub-Saharan
Africa seems to be the region with the highest diversity of Echinococcus, and wild carnivores may play
a more important role in the lifecycles of various species than previously assumed. Still, a number of
issues remain unclear, e.g. possibly diverging parameters of diagnostic tests among the species, different
responses to vaccines and, importantly, possibly required modifications of clinical management due to
differences in pathogenicity.
© 2015 Elsevier B.V. All rights reserved.
1. A summarised history of Echinococcus nomenclature and
taxonomy
1.1. Early period
Echinococcosis of humans and livestock has been known – and
named in various languages – since antiquity. After the introduc-
∗
Corresponding author at: Universität Hohenheim, FG Parasitologie 220 B, Emil-
tion of modern zoological nomenclature in 1758, intended to bring
Wolff-Str. 34, 70599 Stuttgart, Germany. Fax: +49 711 459 22276.
order into the infinite number of poorly defined local names for
E-mail addresses: [email protected] (T. Romig),
animals, naming of what we define today as Echinococcus spp.
[email protected] (D. Ebi), [email protected]
(M. Wassermann). became rather chaotic initially. No less than 85 bi- or trinomial
http://dx.doi.org/10.1016/j.vetpar.2015.07.035
0304-4017/© 2015 Elsevier B.V. All rights reserved.
Please cite this article in press as: Romig, T., et al., Taxonomy and molecular epidemiology of Echinococcus granulosus sensu lato. Vet.
Parasitol. (2015), http://dx.doi.org/10.1016/j.vetpar.2015.07.035
G Model
VETPAR-7722; No. of Pages 9 ARTICLE IN PRESS
2 T. Romig et al. / Veterinary Parasitology xxx (2015) xxx–xxx
Table 1
Synopsis of relevant descriptions of species and subspecies. Agents of cystic echinococcosis (E. granulosus sensu lato) in bold print.
Original name (description) Original description from (stage, host, country) Current name
Hydatigena granulosa Batsch (1796) Metacestode, sheep, Germany Echinococcus granulosus
Taenia multilocularis Leuckart (1863) Metacestode, human, Germany E. multilocularis
T. oligarthra Diesing (1863) Adult, puma, Brazil E. oligarthra
Echinococcus cruzi Brumpt and Joyeux (1924) Metacestode, agouti, Brazil E. oligarthra
E. minimus Cameron (1926) Adult, wolf, Europe E. granulosus
E. longimanubrius Cameron (1926) Adult, African wild dog, South Africa E. granulosus
E. cameroniOrtlepp (1934) Adult, red fox, Britain E. granulosus
a
E. lycaontis Ortlepp (1934) Adult, African wild dog , South Africa E. granulosus
E. felidis Ortlepp (1937) Adult, lion, South Africa E. felidis
b
E. intermediusLopez-Neyra and Soler Planas (1943) Adult, dog, Spain E. granulosus
E. ortleppi Lopez-Neyra and Soler Planas (1943) Adult, dog, South Africa E. ortleppi
E. sibiricensis Rausch and Schiller (1954) Adult, arctic fox, St. Lawrence Isl. E. multilocularis
E. patagonicus Szidat (1960) Adult, Lycalopex, Argentina E. granulosus
c b
E. granulosus canadensisWebster and Cameron (1961) Metacestode , reindeer, Canada E. canadensis
c b
E. granulosus borealis Sweatman and Williams (1963) Metacestode , moose, Canada E. canadensis
d
E. granulosus equinusWilliams and Sweatman (1963) Adult, dog , Britain E. equinus
e
E. granulosus africanus Verster (1965) Adult , div. Canidae, South Africa E. granulosus
E. pampeanus Szidat (1967) Adult, Leopardus colocolo, Argentina E. oligarthra
E. granulosus dusicyontis Blood and Lelijveld (1969) Adult, Lycalopex, Argentina E. granulosus
E. cepanzoiSzidat (1971) Adult, Lycalopex, Argentina E. granulosus
E. vogeli Rausch and Bernstein (1972) Adult, bush dog, Ecuador E. vogeli
E. shiquicus Xiao et al. (2005) Adult, Tibetan fox, China E. shiquicus
E. russicensis Tang et al. (2007) Adult, corsac fox, China E. multilocularis
a
From cyst of sheep origin.
b
Species status under evaluation.
c
Supplemented by adult worms from experimental infections.
d
From cyst of horse origin.
e
From cyst of cattle origin.
Latinised names were published until the end of the 19th century, morphology, e.g. differences in the number of proglottids, rostellar
almost all of them based on metacestodes of various morpholog- hook morphology, number and distribution of testes and position
ical appearance and host origin (Abuladze, 1964). The first valid of the genital pore. In addition to E. granulosus (Batsch, 1786) and E.
name of these was Hydatigena granulosa, given by Batsch in 1786 multilocularis Leuckart, 1863 – whose descriptions were based on
1
and recognisably based on a fertile Echinococcus cyst of sheep ori- metacestodes – Echinococcus oligarthra was described by Diesing
gin from Germany. Shortly after, Rudolphi established the genus (as Taenia) as early as 1863 (later, the metacestode was sepa-
Echinococcus in 1801, the name referring to the small, round, ‘spiny’ rately described under the synonym E. cruzi Brumpt and Joyeux,
protoscolices found in the cysts, and thus created the combina- 1924). Echinococcus minimus and Echinococcus longimanubrius were
tion E. granulosus, which is still in use today. Not recognising the described by Cameron (1926) from a European (Macedonian)
link between metacestodes and adult worms, Rudolphi described wolf and an African wild dog, respectively (Brumpt and Joyeux,
adult Echinococcus from a dog as Taenia cateniformis in 1808. An 1924; Cameron, 1926; Diesing, 1863). Ortlepp added Echinococcus
additional description of adult worms was provided by Beneden in cameroni (for worms from a British fox that Cameron had iden-
1856, as Taenia nana, in ignorance of the fact that three years ear- tified earlier as E. granulosus) and Echinococcus lycaontis (from an
lier the relationship between cysts and adult worms in dogs had African wild dog), followed by Echinococcus felidis from an African
already been proven after independent feeding experiments by von lion (Ortlepp, 1934; Ortlepp, 1937). Echinococcus sibiricensis Rausch
Siebold and Küchenmeister. Eventually, at the end of the 19th cen- and Schiller, 1954 was shortly after description synomymised E.
tury, the common name E. granulosus referred to all stages of the multilocularis (Vogel, 1955; Vogel, 1957). This was followed by
lifecycle, although synonyms like Taenia echinococcus remained in the descriptions of E. intermedius and E. ortleppi (Lopez-Neyra and
use for a long time after. Despite the large number of names that Soler Planas, 1943) from domestic dogs in Spain and South Africa,
had been given to cysts due to their morphological appearances, respectively, and Echinococcus patagonicus (Szidat, 1960), from a
echinococcosis eventually was commonly assumed to be caused wild South American canid (Lycalopex culpaeus) (Lopez-Neyra and
by a single species. Even the metacestode of alveolar echinococco- Soler Planas, 1943; Szidat, 1960). However, in a concise evaluation
sis (described as Echinococcus multilocularis Leuckart, 1863) with of published morphological data, Rausch and Nelson sank most of
its extremely divergent morphology and pathology in humans, and these names as synonyms under E. granulosus, mainly on grounds of
its peculiar geographical restriction, was viewed by the majority of uncertainty about the extent of variability of the diagnostic charac-
authors (the ‘unicists’) as a modification of E. granulosus. The ‘dual- ters used (Rausch and Nelson, 1963). Only two additional species,
ists’ – postulating a different species causing this disease – were in E. multilocularis and E. oligarthra, were considered valid by these
defensive position until the 1950s, when the lifecycle of E. multiloc- authors, while E. felidis and E. patagonicus were given uncertain sta-
ularis was discovered almost simultaneously on St. Lawrence Island tus awaiting further data. A further five species were described later
off Alaska and in central Europe (lit. in Abuladze, 1964; Tappe et al., on, of which E. pampeanus Szidat, 1967 and E. cepanzoi Szidat, 1971
2010). were synonymised with E. oligarthra and E. granulosus, respectively
(Schantz et al., 1976), and E. russicensis Tang et al., 2007 is now
thought to be a variant of E. multilocularis (Nakao et al., 2013a).
1.2. Species
Despite the debate whether echinococcosis might be caused by
one or by two species, a substantial number of additional Echinococ- 1
For the spelling of Echinococcus oligarthra (vs. E. oligarthrus) see Hüttner and
cus species had meanwhile been described based on adult worm Romig, 2009 and Nakaoet al.
Please cite this article in press as: Romig, T., et al., Taxonomy and molecular epidemiology of Echinococcus granulosus sensu lato. Vet.
Parasitol. (2015), http://dx.doi.org/10.1016/j.vetpar.2015.07.035
G Model
VETPAR-7722; No. of Pages 9 ARTICLE IN PRESS
T. Romig et al. / Veterinary Parasitology xxx (2015) xxx–xxx 3
Table 2
were subsequently synonymised with E. oligarthra and E. granulo-
Current concept of Echinococcus taxonomy (agents of cystic echinococcosis in bold).
sus, respectively (Schantz et al., 1976).
Species Genotypes and strains
1.4. Strains and genotypes
Echinococcus granulosus Batsch (1796) G1-3, sheep / buffalo
strains
E. equinus Williams and Sweatman (1963) G4, horse strain At the beginning of the 1980s, there were finally four undis-
E. ortleppi Lopez-Neyra and Soler Planas (1943) G5, cattle strain
puted species (E. granulosus, E. multilocularis, E. oligarthra and E.
E. canadensis Webster and Cameron (1961) G6-7, camel-pig strain G8,
vogeli) (Kumaratilake and Thompson, 1982). It was clear, however,
American‘ cervid strain
that E. granulosus contained a substantial number of variants with
G10, Fennoscandian‘ cervid
strain differences concerning morphology, host specificity, biochemical
E. felidis Ortlepp (1937) lion strain
parameters, developmental biology and geographical distribution.
E. multilocularis Leuckart (1863)
Although the application of the biological species concept by
E. shiquicus Xiao et al. (2005)
Rausch (1967), which had led to the abolishment of all sym-
E. oligarthra Diesing (1863)
E. vogeli Rausch and Bernstein (1972) patric subspecies, had received criticism (Beveridge, 1974), no
attempt was made to resurrect subspecies names for these vari-
ants. Instead, an informal system of intraspecific ‘strains’ was
gradually established. This term was used to describe variants
Only E. vogeli Rausch and Bernstein, 1972 and E. shiquicus Xiao
that differed from each other in characters of epidemiological sig-
et al., 2005 are now considered to be valid species (Nakao et al.,
nificance (Thompson and McManus, 2001). Fully developed, the
2013a) (Table 1) ). Their phylogenetic relationships, based on four
system comprised eleven strains, namely sheep, Tasmanian sheep,
mitochondrial genes (cox1, nad1, cob, rrn), are illustrated in Fig. 1
buffalo, horse, cattle, camel, pig, variant pig (or human-pig), Amer-
ican cervid, Fennoscandian cervid and lion strain. Originally, the
1.3. Subspecies strain system was based on non-genetic characters like host spec-
trum, geography, morphology and aspects of development. From
In addition to species, various subspecies of E. multilocularis and the early 1990s, gene sequence data became increasingly impor-
E. granulosus were described, mostly based on worm morphology. E. tant to define and identify the strains. Important contributions
multilocularis and E. granulosus were subsequently divided into var- to the consolidation of these infraspecific categories were the
ious subspecies, again largely based on morphological characters publications of partial sequences of the mitochondrial cox1 and
of the worms. This started in 1957, when Vogel sank E. sibiri- nad1 genes for seven strains of E. granulosus, plus E. multilocu-
censis, described only three years earlier by Rausch and Schiller, laris, E. vogeli and E. oligarthra (Bowles et al., 1992; Bowles and
as a subspecies under E. multilocularis (a third subspecies E. m. McManus, 1993a). Sequence data corresponded well to other char-
kazakhensis Shul’ts, 1962 was described from metacestodes in acters defining the strains, and led to a genotype ‘nomenclature’
ungulates) (Rausch and Schiller, 1954; Shul’ts, 1962; Vogel, 1957). (G1 to G7), partly replacing the previous strain names. Although
E. g. canadensis Webster and Cameron (1961) was erected due to only a limited number of isolates from biologically or epidemio-
host preference (reindeer) of the cyst stage. Retaining this, Sweat- logically characterised strains were genotyped, the terms ‘strain’
man and Williams added E. g. borealis (moose – canid cycle) and and ‘genotype’ were increasingly treated as synonyms. In addi-
E. g. equinus (horse – dog cycle) in addition to the nominate E. g. tion to the seven genotypes/strains that were initially characterised
granulosus from domestic sheep, cattle and pigs (Sweatman and (G1/sheep strain; G2/Tasmanian sheep strain; G3/buffalo strain;
Williams, 1963; Williams and Sweatmen, 1963). In her major tax- G4/horse strain; G5/cattle strain; G6/camel strain; G7/pig strain),
onomic revision of 1965, Verster retained the subspecies borealis over time three additional taxa were added: the American cervid
and canadensis, resurrected felidis, lycaontis and ortleppi – now strain (G8) (Bowles et al., 1994), a variant pig (or human-pig) strain
as subspecies –, described a new subspecies africanus (from cat- (G9) (Scott et al., 1997), and finally the Fennoscandian cervid strain
tle, sheep and dogs in South Africa), changed the designation of (G10) (Lavikainen et al., 2003). Lack of material for sequencing kept
Sweatman and Williams’ nominate subspecies to E. g. newzealan- the lion strain from being included in the ‘G-system’.
densis, and assigned nominate subspecies status to worm material
from Germany (from where the type species had been described) 1.5. Species, once more
(Verster, 1965). This amounted to eight subspecies of E. granulosus
(the taxa equinus, cameroni, intermedius, longimanubrius, minimus, After two decades of accumulating epidemiological, biochem-
patagonicus and oligarthra were not included in the revision). ical and geographic data on the E. granulosus strains, and
Shortly after, in his largely theoretical treatment of infraspecific cat- the phylogenetic evaluation of increasingly long mitochondrial
egories in Echinococcus, Rausch (1967) refuted the subspecies status and nuclear gene sequences (including complete mitochondrial
for sympatrical forms of E. granulosus, and attributed the mor- genomes), limitations and contradictions of the strain/genotype
phological differences found by previous authors to host-induced system within E. granulosus became apparent and called for a tax-
modifications (Rausch, 1967). A possible exception to this was seen onomic revision of the genus. Major points were (1) the apparent
in the distinction between a ‘northern form’ (or E. g. canaden- paraphyly of E. granulosus (sensu lato) with respect to E. multiloc-
sis), transmitted in wildlife cycles in North America and northern ularis and its sister taxon E. shiquicus (which had in the meantime
Eurasia, and a domestic form (E. g. granulosus), which had been been described from the Tibetan plateau; Xiao et al., 2005, 2006),
globally distributed through human activities. Yet, shortly after and (2) the fact that genetic distances between some genotypes
this consolidation, three additional species were described from (G1-3, G6-7) were in the range of microvariants of the same taxon,
South America: E. pampeanus Szidat (1967) from a wild cat species, whereas others (G4, G5) were only distantly related. After a first
E. cepanzoi Szidat (1971) (as a new name for the subspecies E. g. proposition to subdivide E. granulosus into four species (Thompson
dusicyontis Blood and Lelijveld, 1969) from a wild South Ameri- et al., 1995), E. granulosus equinus Williams and Sweatman, 1963
can canid (Lycalopex sp.), and finally E. vogeli Rausch and Bernstein was finally elevated to species rank (for the horse strain, genotype
(1972); from the bush dog (Speothos venaticus) (Blood and Lelijveld, G4), and E. ortleppi Lopez-Neyra and Soler Planas, 1943 was rein-
1969; Rausch and Bernstein, 1972; Szidat, 1967; Szidat, 1971). Of stated (for the cattle strain, genotype G5; Thompson and McManus,
these, only E. vogeli survived the test of time, while the first two 2002). This left the name E. granulosus (Batsch, 1786) for the
Please cite this article in press as: Romig, T., et al., Taxonomy and molecular epidemiology of Echinococcus granulosus sensu lato. Vet.
Parasitol. (2015), http://dx.doi.org/10.1016/j.vetpar.2015.07.035
G Model
VETPAR-7722; No. of Pages 9 ARTICLE IN PRESS
4 T. Romig et al. / Veterinary Parasitology xxx (2015) xxx–xxx
Fig. 1. Cladogram of Echinococcus spp. obtained through Maximum Likelihood anal-
yses of 5170 nucleotides of the mitochondrial cox1, nad1, rrn and cob gene. Modified
from Hüttner et al. (2008). Fig. 2. E. granulosus sensu stricto: haplotype network of the complete mitochon-
drial cox1 gene (1608 bp). The network consists of 137 haplotypes based on 304
analysed isolates from Europe (16), western Asia (140), southern/eastern Asia (40),
genotypes G1 to G3 (sheep, Tasmanian sheep and buffalo strains), Africa (78) and South America (30) (authors, unpublished data and Genbank entries).
while the allocation of the camel, pig, cervid and lion strains was The network includes haplotypes published by Konyaev et al. (2013) and Yanagida
et al. (2012). For comparison, one isolate of E. felidis is included (Genbank accession
left unresolved until, five years later and based on the compari-
no. AB732958). The network was constructed using TCS 1.21 (Clement et al., 2000),
son of complete mitochondrial genomes, E. granulosus canadensis
with fixed connection limit at 130 steps.
Webster and Cameron, 1961 was given species status, now includ-
Circle sizes are not proportional to the haplotype frequencies. Large central cir-
ing several closely related genotypes (G6 to G10) (Nakao et al., cle: represents the most common, globally distributed haplotype (e.g. accession no.
JQ250806). Medium circles: identified haplotypes. Small circles: hypothetical inter-
2007). Finally, E. felidis Ortlepp, 1937; could be resurrected from
mediate haplotypes (not identified in this panel). Rectangle: represents 28 different
synonymy based on mitochondrial sequences obtained from pre-
haplotypes with one basepair difference to the central haplotype. Black circles: hap-
served adult worm material that had been determined by Verster in
lotypes that contain the G1 sequence of Bowles et al. (1992) (21 of 28 haplotypes
South Africa (Hüttner et al., 2008). In the current state of this (ongo- of the rectangle belong to this type). Dark grey circle: haplotypes that contain the
ing) taxonomic reshuffle, E. granulosus in its previous sense (or G2 sequence of Bowles et al. (1992). Light grey circles: haplotypes that contain the
G3 sequence of Bowles et al. (1992). White circles: not described in the G-system of
sensu lato) is split into five species (granulosus s.s., felidis, equinus,
genotypes.
ortleppi, canadensis), in addition to the agents of alveolar and poly-
cystic echinococcosis (E. multilocularis, shiquicus, oligarthra, vogeli)
(Table 2.
nad1 sequences published in the 1990s, the use of the G num-
bers has become increasingly misleading (see also discussion on
2. Species accounts and molecular epidemiology this by Nakao et al., 2013a). A substantial number of different gene